We present the WFD (Wheat Fungi Diseases) dataset of 2414 wheat images. Each image was manually marked by experts. The images were classified into healthy plants and by the presence of lesions on them with the following types of fungal diseases: leaf rust, powdery mildew, septoria, stem rust, yellow rust. Additionally, the images were classified according to the stage of plant development: whether the plant is a seedling (seedling tag) or not (no such tag). Note that images could be tagged with several tags at the same time. In total, our dataset contained 3147 tags. The total size of the images is 4.84GB. The average image resolution in this dataset is ~ 5MP. Based on this dataset, we trained the EfficientNet-B0 model, which, from a digital image of wheat plants, determines the stage of development and type of disease with an accuracy of 94.2%.
If you use the dataset for your paper, please cite: Genaev, M.; Skolotneva, E.; Gultyaeva, E.; Orlova, E.; Bechtold, N.; Afonnikov, D. Image-Based Wheat Fungi Diseases Identification by Deep Learning. Plants 2021 (doi: 10.3390/plants10081500).
Prior to downloading the data, please ensure that you read the terms and conditions.
To download the data please fill in your contact/affiliation information in this form. Having your contacts we can reach out to you and share updates.
We hope the data are useful for your research.